Data 360 object model and APIs

Data Cloud was renamed Data 360 on October 14, 2025. Metadata types, scratch-org features and many pages still say Data Cloud or CDP.

The layers

Layer	Suffix	What it is
Data stream → data lake object (DLO)	`__dll`	Raw ingested rows, one DLO per stream
Data model object (DMO)	`__dlm`	Harmonised data: DLO fields mapped to the Customer 360 model. Standard DMOs carry `ssot__` in object and field names
Unified DMO	`__dlm`	The golden record that identity resolution writes, plus a link DMO joining each source record to it
Calculated insight object (CIO)	`__cio`	A SQL-defined metric (lifetime value, last order) per profile or account
Search index chunk and index objects	`__dlm`	Created by a search index for retrieval (`reference/data360/search-indexes-and-retrievers.md`)

The revenue DMOs

DMO	API name	Key fields for revenue
Individual	`ssot__Individual__dlm`	`ssot__Id__c`, names, birth date
Contact Point Email / Phone / Address	`ssot__ContactPointEmail__dlm` and so on	Email, phone and address, each linked to the party
Account / Account Contact	`ssot__Account__dlm`, `ssot__AccountContact__dlm`	Account name, number, parent, type; contact-to-account link
Sales Order	`ssot__SalesOrder__dlm`	`ssot__SoldToCustomerId__c`, `ssot__BillToAccountId__c`, `ssot__GrandTotalAmount__c`, `ssot__PurchaseOrderDate__c`, `ssot__OrderStartDate__c`, `ssot__SalesOrderStatusId__c`, `ssot__CurrencyCode__c`
Sales Order Product	`ssot__SalesOrderProduct__dlm`	Line items; many to one with Sales Order
Opportunity, Case, Master Product	`ssot__Opportunity__dlm` and so on	Pipeline, service history, catalogue
Unified Individual and its link	`UnifiedIndividual__dlm`, `IndividualIdentityLink__dlm` (names in Salesforce's Trailhead SQL)	The golden record and its source links

Don't hard-code unified names. When an org has more than one ruleset, the ruleset ID is appended to the unified and link object names; Salesforce's workshop shows "Unified Individual ccid". List the DMOs and pick them up by name pattern: GET /services/data/v66.0/ssot/data-model-objects returns dataModelObject[], and …/data-model-objects/{name} returns its fields[].

Two ways to call Data 360

	Connect REST API	Tenant APIs
Base	`https://<My Domain>/services/data/v66.0/ssot/...`	The `instance_url` returned by the token exchange, for example `/api/v1/metadata`
Token	The normal Salesforce access token (the CLI's session works)	Exchange it at `https://<My Domain>/services/a360/token` with `grant_type=urn:salesforce:grant-type:external:cdp`, `subject_token=<access token>`, `subject_token_type=urn:ietf:params:oauth:token-type:access_token` and optionally `dataspace=<name>`
Scopes	`api`; Data 360 setup operations also need `cdp_api` (per the community plugin's guide)	`cdp_query_api` (SQL), `cdp_profile_api` (profiles), `cdp_ingest_api` (Ingestion API), plus `refresh_token`
Use for	Setup and lifecycle (streams, mappings, identity resolution, insights, search indexes, data kits) and SQL through `query-sql`	Ingestion API, Profile API, direct Query API v2 and v3

Salesforce calls the Connect REST API "the preferred platform for customers to construct solutions on top of Data 360". The user also needs Data 360 permission sets for the objects it touches.

Query Connect API

POST /services/data/v66.0/ssot/query-sql?dataspace=default
{ "sql": "SELECT COUNT(*) AS n FROM ssot__SalesOrder__dlm WHERE ssot__PurchaseOrderDate__c >= :since",
  "sqlParameters": [ { "type": "date", "name": "since", "value": "2026-01-01" } ] }

The response holds:
returnedRows;
metadata[], with each column's name, type and nullable;
data[][];
status, with queryId, completionStatus, rowCount, chunkCount, progress and expirationTime.
Large results:
Poll GET …/query-sql/{queryId}?waitTimeMs=5000 until finished, then page with GET …/query-sql/{queryId}/rows?rowLimit=2000&offset=2000&omitSchema=true.
Cancel with DELETE …/query-sql/{queryId}.
Status values: Salesforce's post shows completionStatus both as Finished and as FINISHED. Compare case-insensitively.
Retention and caching: results live for 24 hours, and identical queries may be served from cache for 24 hours.
Queries read DMOs, DLOs and CIOs by API name. Quote mixed-case names, as in "ssot__Individual__dlm", if your SQL client lower-cases them.

From TypeScript, conn.request({ method: "POST", url: "/services/data/v66.0/ssot/query-sql", body }) with an @salesforce/core connection is enough. The pipeline guide's code does this: guides/data360/automate-the-golden-record-pipeline.md.